Since we didn't check and insist that args.pad must be zero for MMAP_GTT
historically, we cannot insert a check now as old userspace may be
feeding in garbage. As such the lack of check is enshrined into the ABI,
so add a comment to remind us we cannot add the check later.

Signed-off-by: Chris Wilson <[email protected]>
Cc: Joonas Lahtinen <[email protected]>
---
 drivers/gpu/drm/i915/gem/i915_gem_mman.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c 
b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
index 3a3f30bc8ac7..15faf95d74ca 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
@@ -620,6 +620,15 @@ i915_gem_mmap_offset_ioctl(struct drm_device *dev, void 
*data,
        enum i915_mmap_type type;
        int err;
 
+       /*
+        * Historically we failed to check args.pad and args.offset
+        * and so we cannot those fields for user input and we cannot
+        * add -EINVAL for them as the ABI is fixed, i.e. old userspace
+        * may be feeding in garbage in those fields.
+        *
+        * if (args->pad) return -EINVAL; is verbotten!
+        */
+
        err = i915_user_extensions(u64_to_user_ptr(args->extensions),
                                   NULL, 0, NULL);
        if (err)
-- 
2.24.0

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Reply via email to